Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Ali Ataya

Pulmonary Vascular Med · UF

View

Dr. Naftali Kaminski

Interstitial Lung Disease · Yale College

View

Dr. Auston Stiefer

Interstitial Lung Disease · CU Anschutz Medical Campus

View

Dr. Justin Zhang

Critical Care · UCLA

View

Dr. Panagis Galiatsatos

Critical Care · Johns Hopkins

View

Dr. MeiLan Han

Asthma & COPD · UMHS

View

Dr. Kevin Chung

Critical Care · SeaStar Medical

View

Dr. Diego J. Maselli

Asthma & COPD · UTHSCSA

View

Dr. Amos E. Dodi

Critical Care · MMC

View

Dr. Imre Noth

Interstitial Lung Disease · University of Virginia

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Elevated Pulmonary Artery Wedge Pressure in Group 1 Pulmonary Hypertension.

    Feb 13, 2026DOI: 10.1161/circulationaha.125.077606Journal ArticleAHA

    Written by: Yogesh N V ReddyRobert P FrantzWilliam R MirandaAneesh K AsokanRevati VarmaFranz RischardPaul M HassounAnna R HemnesEvelyn HornJane A LeopoldErika B RosenzweigNicholas S HillSerpil C ErzurumGerald J BeckJohn BarnardJ Emanuel FinetDeborah KwonStephen C MathaiMonica MukherjeeW H Wilson TangK Sreekumaran NairBarry A Borlaug

    Find these authors on Roon

    YR

    Dr. Yogesh Reddy

    Advanced Heart Failure · Mayo Clinic College of Medicine

    AH

    Dr. Anna Hemnes

    Pulmonary Vascular Med · Vanderbilt

    Introduction

    Patients with group 1 pulmonary hypertension (PH) have a primary pulmonary vasculopathy that leads to elevation in pulmonary vascular resistance (PVR). Since left heart disease and heart failure with preserved ejection fraction (HFpEF) can also cause elevation in PVR, guidelines have recommended a resting pulmonary artery wedge pressure (PAWP) threshold of 15 mm Hg to diagnose left heart dysfunction based on reported PAWP values from healthy controls.1 The cause of PH and associated PVR elevation is therefore currently dichotomized by resting PAWP, where patients with a PAWP≤15 mm Hg are considered to have group 1 PH [after excluding lung (group 3) or thromboembolic (group 4) related PVR elevation), whereas those with PAWP>15 mm Hg are diagnosed with left heart related PVR elevation i.e. combined pre and post capillary (Cpc) PH HFpEF.2,3As with any binary dichotomization process, utilization of a single resting PAWP cutoff may introduce disease misclassification.4–6 A subset of patients are diagnosed and treated as group 1 PH despite PAWP>15 mm Hg,7–9 but such patients are routinely excluded from clinical trials that mandate a resting PAWP≤15 mm Hg.10 We hypothesized that group 1 PH patients with PAWP>15 mm Hg would have precapillary pulmonary vascular pathology more similar to traditional group 1 PH, and distinct from Cpc PH HFpEF. To test our hypothesis, we used the multicenter PVDOMICS dataset where all individuals underwent hemodynamic provocation, manual re-adjudication of all recorded pressures at end-expiration, unbiased metabolomic evaluation, and integrated PH specialist adjudication of the diagnosis of group 1 PH with normal PAWP, group 1 PH despite high PAWP, or Cpc PH HFpEF. We also independently tested the prevalence and pathophysiology of resting PAWP elevation in group 1 PH among two independent validation cohorts undergoing exercise RHC, and a third validation cohort undergoing resting RHC, with all patients undergoing probabilistic adjudication of group 1 PH or Cpc PH HFpEF by PH specialists.

    Methods

    The PVDOMICS clinical research network is a National Heart, Lung, and Blood Institute–funded, prospective, longitudinal cohort study (NCT02980887) that enrolled participants from November 30, 2016, to October 18, 2019 at seven centers across the US.11 The protocol was approved at each institution by the local Institutional Review Board, and informed consent was obtained from all participants.All participants in PVDOMICS underwent formal adjudication of their ‘primary’ World Symposium on Pulmonary Hypertension group using integrated probabilistic assessment of all available diagnostic information. Initial group assignments were by the recruiting investigators who were practicing PH physicians at each site. To avoid tautological bias inherent to classification by resting PAWP thresholds alone, patients were allowed to be categorized as group 1 PH despite a PAWP>15 mmHg. However, additional adjudication of this subset of patients was performed by an independent PVDOMICS adjudication committee.
    The final PVDOMICS adjudicated PH diagnosis was based on the clinical, imaging and hemodynamic profile. This included resting and provoked hemodynamics, severity of PAWP and PAWP v wave elevation relative to mean pulmonary artery (mPA) pressure and right atrial (RA) pressure, severity of pulmonary vascular dysfunction, prior history and clinical response to PH vasodilator therapy, and comprehensive cardiac imaging. The full consensus criteria utilized are detailed in the online supplement and summarized in Table 1.
    Group 1 PH was therefore defined as mPA>20 mm Hg with PVDOMICS adjudicated ‘primary’ group 1 PH. For the main analysis, group 1 PH patients were dichotomized based on resting PAWP, grouping those with a normal PAWP≤15 mm Hg (group 1 PH + normal PAWP) and those with an abnormal PAWP>15 mm Hg (group 1 PH + high PAWP).
    As an additional reference population, we also included patients with Cpc PH due to HFpEF. We focused on the HFpEF population rather than group 2 PH in general, as these patients are most at risk for misdiagnosis/overlap with group 1 PH.8 Included patients with Cpc PH HFpEF met resting hemodynamic criteria (mPA>20 mm Hg, PAWP>15 mm Hg and PVR>2 Wood units) and were independently adjudicated by PVDOMICS investigators as group 2 PH related to HFpEF. As an agnostic, adjudication independent estimate of the likelihood of HFpEF, we also quantified pretest probability for HFpEF using the validated HFpEF-ABA algorithm based only on clinical characteristics.12 Healthy asymptomatic controls recruited to PVDOMICS formed another reference population for normative values of imaging measures.(Figure S1)
    Participants underwent detailed baseline evaluation including imaging, functional and hemodynamic characterization as previously described (details in online supplement). Right heart catheterization (RHC) was performed at rest in the supine position with manual re-adjudication of pressure measurements at end-expiration as previously described.13 Patients then underwent repeat hemodynamic assessment during i) 100% oxygen challenge for 5 minutes and ii) inhaled nitric oxide (NO) challenge at 40 ppm for 5 minutes added to 100% oxygen with assessment of PAWP, mPA and cardiac output (CO) by thermodilution in both phases. The PAWP v wave was manually measured at end-expiration as the highest point on the PAWP tracing timed around the EKG T wave. Hemodynamic stress for left heart disease was performed based on center availability either through i) fluid challenge with 500 mL of 0.9% saline over 10-minutes or ii) invasive exercise hemodynamics to peak exhaustion in the supine or upright position based on clinical practice patterns at recruiting centers.
    The contribution of myocardial stiffness of the left ventricle to PAWP elevation was estimated using cardiac MRI and RHC derived stiffness constant β (increases with greater LV stiffness), and V30 (decreases with greater LV stiffness). Left atrial (LA) operating compliance and other LA function indices were calculated as previously described (details in online supplement).14,15
    All PVDOMICS participants underwent fasting venous plasma collection for untargeted metabolomics as previously described.16(details in online supplement)
    Validation cohort 1 included consecutive patients undergoing supine exercise RHC at the Mayo Clinic between 2006–2024. PH diagnosis was adjudicated independently by YNR and WRM (details, online supplement). Catheterizations included simultaneous gas exchange measures, high-fidelity micromanometers, and exercise hemodynamics as previously described.17,18 All tracings were manually remeasured at end-expiration by a single cardiologist with experience in hemodynamic interpretation (YNR) while blinded to clinical information. From simultaneous RA and PAWP measures, the contribution of pericardial restraint to PAWP elevation was estimated by increasing RA/PAWP ratio and a decrease in LV transmural pressure (LVTMP= PAWP-RA)(additional details in online supplement).19–23
    Similar to the PVDOMICS cohort, the adjudication of group 1 PH was made probabilistically using all available clinical information and detailed evaluation of hemodynamic tracings at rest and exercise. Adjudication of group 1 PH with resting PAWP>15 mm Hg was allowed if the final diagnosis was most consistent with group 1 PH on probabilistic adjudication. As in the other cohorts, prior use of vasodilators alone was insufficient to quality for a diagnosis of group 1 PH, and the ultimate diagnosis considered all available clinical and hemodynamic factors. As a comparator group, we also included patients with definite Cpc PH HFpEF at rest (mPA>20mmHg, PVR>2Wood units, rest PAWP>15mmHg), along with persistent elevation in PAWP during exercise≥25mmHg.
    Validation cohort 2 included patients from the Mayo clinic with adjudicated group 1 PH who underwent exercise RHC between 2024–2025 with the exact same criteria as above (adjudication by YNR and WRM).
    Validation cohort 3 included patients between 2000–2020 from the Mayo Clinic who underwent resting RHC with prior or new prescription for PH medications following RHC. From this cohort of PH therapy treated patients, the historical PH provider adjudicated diagnosis was abstracted by detailed chart review by an investigator (RV), followed by independent readjudication by YNR of all cases and hemodynamic tracings as most consistent with either i) group 1 PH or ii) Cpc PH HFpEF despite use of PH therapy.
    The data that support the findings of this study are available from the corresponding author on reasonable request. Differences across the groups were compared using ANOVA, Kruskal-Wallis test and chi-square test as appropriate. Individual group comparisons were performed by Tukey and Steel-Dwass test as appropriate in the setting of multiple comparisons. Missing values were not imputed for descriptive analyses. To account for the correlated nature of repeated hemodynamic measures, we utilized a mixed model approach with a compound symmetry covariance structure using participant ID as a random effect with group, phase and their interaction as fixed effects. Cpc PH HFpEF and Group 1 PH patients stratified by resting PAWP were followed for all-cause mortality using the log rank test from time of study enrollment. Cox proportional hazard models were used to determine the hazard ratio for death between groups.
    Details of metabolomic data preparation for analyses are included in the online supplement. Following metabolite data normalization, we performed univariate two-sample t-tests to identify significantly changed metabolites comparing group 1 PH + high PAWP vs group 1 PH + normal PAWP, and separately compared the metabolome between group 1 PH + high PAWP vs Cpc PH HFpEF. The Benjamini-Hochberg method was used to correct the false discovery rate(FDR) for the 954 metabolites tested with an adjusted p<0.05 considered statistically significant. Given the importance of obesity to metabolome changes24–26, we also present metabolite differences adjusted for baseline BMI using general linear models with covariate adjustment(Limma package). Pathway enrichment analysis was performed using metabolite set enrichment overrepresentation analysis. We used the list of all measured metabolites with KEGG IDs as a reference metabolome for enrichment tests. We selected the RaMP-DB-based pathway database for enrichment analysis, integrating pathways from KEGG, HMDB, Reactome, and Wiki Pathways.27 We evaluated all significant upregulated or downregulated metabolites separately for pathway analysis. The current report adhered to STROBE reporting guidelines. Statistical analyses were performed using MetaboAnalyst 6.0 software and JMP, version 17.1.0(JMP Statistical Discovery LLC).

    Results

    Out of 325 patients with adjudicated group 1 PH (mean age 53.0±14.3, female 73%[n=236]), 85% (n=277) had normal resting PAWP and 15% (n=48) had high resting PAWP(Figure S1). Compared with both group 1 PH subgroups, patients with Cpc PH HFpEF (n=75) were older (mean age 68.9±11.4, 64% female[n=23]) with greater hypertension and diabetes but also had the highest natriuretic peptide levels and diuretic usage, worst renal function, greatest inflammation (c-reactive protein) and lowest hemoglobin (Table 2). There was progressive increase in atrial fibrillation burden and obesity from group 1 PH + normal PAWP to group 1 PH + high PAWP to Cpc PH HFpEF, with an associated increase in pretest probability for HFpEF based on the HFpEF-ABA algorithm (Figure S2). Most patients with group 1 PH had known treated PH prior to enrollment with high rate of baseline pulmonary vasodilator use (96%[n=46]) in group 1 PH + high PAWP, and longer duration of known PH prior to enrollment.
    Patients with Cpc PH HFpEF had the largest LV and LA volumes, greater LV diastolic dysfunction (lower lateral e’ and higher E/e’) and greater LA dysfunction (lower LA EF and expansion index) compared to both group 1 PH subgroups (Table 3)(Table S1)(Figure S3). Patients with Cpc PH HFpEF also had larger RA volumes and greater impairment in RA ejection fraction, consistent with more biatrial myopathy in this group (Table S1).
    Group 1 PH + high PAWP had the largest right ventricular size and the poorest RV-PA coupling compared to other groups, while pulmonary artery enlargement was highest in both group 1 PH subgroups. Group 1 PH + normal PAWP had the lowest LV mass and highest LV longitudinal strain (Table 3).
    To contextualize cardiac remodeling observed in group 1 PH + high PAWP, we next contrasted this group to asymptomatic healthy controls (n=96) recruited to PVDOMICS. Group 1 PH + high PAWP was associated with greater obesity, modestly higher hypertension, worse renal function and higher pre-test probability for HFpEF (Table S2). As expected, right heart remodeling and function indices by echocardiogram and MRI were all abnormal relative to healthy controls. Notably, measures of left heart performance and remodeling were also abnormal in group 1 PH + high PAWP compared to healthy controls including LV mass, LV longitudinal strain, LA size, LA EF, LA expansion index, e’ velocity and E/e’ (Figure S3)(Table S3).
    Patients with Cpc PH HFpEF had the greatest symptom burden (higher NYHA class, worse QOL), and poorest exercise tolerance (peak VO2 and 6-minute walk distance) (Table S1). Functional impairment was largely comparable in the two group 1 PH subgroups, regardless of resting PAWP.
    By definition, group 1 PH + high PAWP patients had higher PAWP compared to group 1 PH + normal PAWP, which was comparable to those with Cpc PH HFpEF. In contrast to the mean PAWP, there was a progressive increase in the amplitude of the PAWP v wave across groups, with the Cpc PH HFpEF group having the greatest v wave indicating poorest LA compliance (Figure 1)(Table 4). Elevation in PVR was greatest in both group 1 PH subgroups compared to Cpc PH HFpEF. As a consequence of the combined elevation in resting PAWP and PVR, patients with group 1 PH + high PAWP had the highest resting mPA pressure compared to other groups(Table 4). There was a positive correlation between PAWP and mPA pressure in group 1 PH overall (r=+0.39, p<0.0001), with directionally similar correlation when restricted to group 1 PH + high PAWP (r=+0.27, p=0.06). Estimated LV stiffness was lowest in the group 1 PH + normal PAWP group, vs. comparable results in group 1 PH + high PAWP and Cpc PH HFpEF (Figure S2).
    Following 100% O2 and inhaled NO, there was a greater reduction in PAWP in group 1 PH + high PAWP compared to Cpc PH HFpEF and group 1 PH + normal PAWP, but there remained a progressive gradient of PAWP abnormalities across the 3 groups. (Table 4)(Table S4)(Figure 1) Other group differences in hemodynamics including larger PAWP v waves in Cpc PH HFpEF persisted from baseline through the O2, NO and fluid challenge phases (Table 4)(Table S4)(Figure 1).
    Considering repeated measurements across all phases, there were significant group differences in the change with provocation for PAWP, PAWP v wave, Δ v wave height and mPA pressure. Baseline group differences in PVR persisted across repeated measurements (Figure 1). During exercise PAWP increased progressively across the three groups, while exercise PVR was highest in the two group 1 PH groups (Table S4). Patients with Cpc PH HFpEF had the lowest peak VO2 and most severe dyspnea with exercise, while the two group 1 PH subgroups had the greatest ventilatory inefficiency (Table S4).
    Metabolomic analyses revealed little difference between group 1 PH with high vs normal PAWP, with only 24 metabolites differing after FDR correction, with modest effect size, and none remaining significant after adjusting for BMI (Figure 2) (online supplement). In contrast, there were large differences in the metabolome between group 1 PH + high PAWP vs Cpc PH HFpEF, with 100 metabolites different after FDR correction, of which 91 remained significant after BMI adjustment (online supplement).
    Compared with group 1 PH + high PAWP, Cpc PH HFpEF was associated with greater deficiency of amino acids associated with metabolic dysfunction (including glycine and serine), nitric oxide signaling (arginine) and cellular senescence (tryptophan) among others (Full list of altered metabolites in the online supplement). Pathway enrichment analysis showed that compared with Group 1 PH + high PAWP, downregulated metabolites in the Cpc PH HFpEF group were mainly involved in amino acid metabolism-related pathways (Figure S4). Notably, 15 out of the 20 proteogenic amino acids (arginine, asparagine, aspartate, cysteine, glutamine, glycine, histidine, isoleucine, lysine, methionine, phenylalanine, proline, serine, threonine, and tryptophan) were decreased in the Cpc PH HFpEF group (online supplement). There were also elevated levels of amino acid catabolism-related metabolites, such as urea, N-acetylcitrulline, and homocitrulline suggesting increased liver degradation through the urea cycle in Cpc PH HFpEF contributing to amino acid deficiency distinct from group 1 PH + high PAWP.28
    Over a median follow-up of 4.7 years [IQR 4.0–6.0] there were 107 deaths in the group 1 PH and Cpc PH HFpEF cohorts. Patients with Cpc PH HFpEF had worse unadjusted survival (log-rank p=0.01)(Figure S5). However, after adjusting for baseline differences in age, NYHA functional class, six minute walk distance and NT-proBNP, there were no significant group differences in survival [Cpc PH HFpEF vs group 1 PH + normal PAWP HR 0.56 (95%CI 0.30–1.05), p=0.07; Cpc PH HFpEF vs group 1 PH + high PAWP HR 0.49 (95%CI 0.21–1.14), p=0.10]
    Validation cohort 1 included 402 patients with PH (31%[n=124] group 1 PH, 69%[n=278] Cpc PH). Out of the 124 patients with group 1 PH, 18%(n=22) had blinded end-expiratory PAWP at rest>15 mm Hg, with comparably severe PVR elevation across both group 1 PH subgroups. Most group 1 patients represented incident new diagnosis of PH in this cohort (79%[n=81] with group 1 PH + normal PAWP vs 68%[n=15] with group 1 PH + high PAWP). Patients with Cpc PH HFpEF had greater age, BMI, hypertension and atrial fibrillation burden compared to group 1 PH, and those with group 1 PH + normal PAWP had the most normal left heart remodeling/function by echocardiography (Table S5).
    Like the PVDOMICS cohort, exercise PAWP and v wave abnormalities were of intermediate severity in group 1 PH + high PAWP compared to the other two groups (group p<0.0001, interaction p<0.0001) (Table 5) (Figure 3)(Figure S6). Patients with group 1 PH + high PAWP had a greater rise in RA/PAWP ratio and decrease in LVTMP without prominent PAWP v waves compared with Cpc PH HFpEF, supporting contribution from enhanced pericardial interaction to the observed exercise PAWP elevation (Figure 3). However, the elevation in exercise PAWP in group 1 PH + high PAWP was not merely related to pericardial restraint alone, as exercise PAWP remained normal in group 1 PH + normal PAWP despite similar pericardial interaction during exercise with comparably low LVTMP and high RA/PAWP, supporting an additional contribution from left heart diastolic dysfunction to the PAWP abnormalities in group 1 PH + high PAWP (Figure 3).
    Validation cohort 2 included 55 patients with adjudicated group 1 PH of whom 22% (n=12) had high PAWP at rest. Group 1 PH + high PAWP remained associated with greater obesity, with comparable prevalence of previously diagnosed group 1 PH with prior use of vasodilators (51%[n=22] with normal PAWP vs 41%[n=5] with high PAWP, p=0.56) (Table S6). Similar to other cohorts, PAWP during exercise was higher in group 1 PH + high PAWP (p<0.0008) compared with group 1 PH + normal PAWP without a prominent rise in PAWP v waves (p=0.73)(Table S7).
    In validation cohort 3 (n=787), there were 665 adjudicated cases of group 1 PH of whom 19% (n=126) had a high PAWP at rest, and 122 adjudicated cases of group 2 PH, of whom 18% (n=22/122) had a normal PAWP at rest. Group 1 PH + high PAWP had greater obesity compared to group 1 PH + normal PAWP, with Cpc PH HFpEF patients having older age, greater atrial fibrillation and more abnormal left heart remodeling by echocardiography. Consistent with other cohorts, there was a progressive gradient of PAWP abnormalities observed across the 3 groups, with highest PA pressure in group 1 PH + high PAWP.(Table S8)
    In the PVDOMICS cohort, adjustment for baseline comorbidity differences (age, hypertension, diabetes, sleep apnea, BMI) and PH treatment history (number of years of PH prior to enrollment, prior PH medication use) still resulted in persistent differences in PAWP and LV remodeling in group 1 PH + high PAWP, suggesting that comorbidity differences and prior PH therapy alone did not explain group differences in left heart hemodynamics (Table S9). Consistent with this, the prevalence of high PAWP in group 1 PH was comparable among prevalent and incident cases in validation cohorts 1+2 [21.8%(n=12) vs 17.7%(n=22), p=0.52], while being slightly more common among prevalent compared to incident cases in validation cohort 3 [26.6%(n=57) vs 15.3%(n=69), p=0.0006]. Among patients who underwent repeat RHC in validation cohorts 1+2 (n=65) there was greater reduction with PH therapy in PAWP, RA pressure and PA pressure in group 1 PH + high PAWP compared to group 1 PH + normal PAWP, with comparable improvement in PVR, cardiac output and NT-proBNP in both groups. These differences in longitudinal hemodynamic response were replicated in those from validation cohort 3 with repeat RHC assessment (n=328) (Table S10).
    Among the Mayo Clinic validation cohort 1 (n=402), there was good agreement between two independent adjudicators (YNR, WRM) of the diagnosis of group 1 PH or Cpc PH HFpEF when considering all available information including hemodynamics in a probabilistic manner (Kappa coefficient 0.89 [95% CI 0.84–0.94], p<0.0001). Among the subset of PVDOMICS participants from the Mayo Clinic where independent readjudication was possible (n=54) (RPF, YNR), there was also good agreement in the diagnosis of group 1 PH or Cpc PH HFpEF (Kappa coefficient 0.92 [95% CI 0.82–1.00], p<0.0001). Among participants initially adjudicated to have group 1 PH, the percent agreement for diagnosis of group 1 PH between two independent reviewers was 93.9%, 86.3%, 96.4% and 84.5% in adjudicated patients from the primary cohort and validation cohorts 1–3 respectively.

    Discussion

    In this study, we report the first comprehensive phenotypic characterization of prospectively adjudicated patients with group 1 PH stratified by resting PAWP. We identify that when the diagnosis of group 1 PH is established by PH experts, around 1 out of 5 patients have adjudicated PAWP elevation at rest that would normally be indicative of group 2 PH. Elevation in PAWP in this group persisted across repeated provocative measurements, and findings were replicated in three independent validation cohorts. Although the identified group 1 PH + high PAWP patients would fulfill current diagnostic criteria for Cpc PH due to resting elevation in PAWP and PVR3, their clinical characteristics, prior pulmonary vasodilator use, and metabolomic profile were more comparable to traditional group 1 PH with normal PAWP. However, there were also key differences including greater obesity, higher pretest probability for HFpEF, and subclinical abnormalities in LV remodeling and diastolic dysfunction by multiple independent measures. The observed left heart changes in group 1 PH + high PAWP were pathologic relative to healthy controls, but of intermediate severity between true Cpc PH HFpEF and more traditionally defined group 1 PH.
    Despite some similarities in LV abnormalities between Cpc PH HFpEF and group 1 PH + high PAWP, there were also notable differences including less severe pulmonary vascular disease, worse LA compliance and higher PAWP v wave response during atrial filling in Cpc PH. The metabolome differed between Cpc PH HFpEF and Group 1 PH + high PAWP, with greater evidence of amino acid deficiency in Cpc PH HFpEF, further supporting key biological differences. Exercise PAWP response in group 1 PH + high PAWP was intermediately abnormal relative to the other two groups, with evidence of enhanced ventricular interaction additionally contributing to abnormal PAWP elevation during exercise. Collectively, these data support the idea that group 1 PH + high PAWP is pathophysiologically distinct from Cpc PH HFpEF, and although these patients appear more closely aligned with a diagnosis of group 1 PH, the unique pathophysiology inherent to also having PAWP elevation suggests they are also distinct from traditionally-defined group 1 PH, calling for further study in trials focused on this subgroup of patients.
    The present data support the hypothesis that patients with group 1 PH + high PAWP truly have precapillary pulmonary vascular pathology. While there is superimposed LV diastolic dysfunction, this is likely not the primary basis for the observed severe PVR elevation. Although recent guidelines and position statements have endorsed the need for probabilistic evaluation of the diagnosis of Group 1 PH versus Cpc PH HFpEF, without excessive reliance on resting PAWP thresholds alone to dichotomize the diagnosis,4,5 the identified group 1 PH subset with PAWP>15 mm Hg has been systematically excluded from clinical trials, so there is inadequate evidence to guide therapy in this subset.
    Post-hoc analyses of randomized trials have suggested no interaction based on baseline PAWP,10 although the systematic exclusion of those with resting PAWP>15 mm Hg limits the generalizability of these findings. Observational studies have suggested worse therapeutic response in those with adjudicated group 1 PH who develop a higher resting PAWP after vasodilator therapy initiation.29 The observation of high baseline vasodilator use among group 1 PH + high PAWP patients in this study might support the hypothesis that chronic pulmonary vasodilator use and improved LV filling may have unmasked latent LV diastolic dysfunction in some patients with group 1 PH. Although this may contribute in some cases, the generally comparable prevalence of PAWP elevation even in incident group 1 PH argues against this being the only explanation for the observed PAWP elevation. Although the mechanisms of PAWP elevation require further study, it seems plausible that a subset of these patients with prevalent long standing group 1 PH may develop independent acquired left heart dysfunction from HFpEF with associated risk factors of obesity and aging. It is also possible that there may be as yet unidentified common risk factors that may contribute to both PAH and HFpEF, and predispose to disproportionate pulmonary vascular remodeling with even modest left heart pathology in a subset of patients.
    Although likely contributory, the presence of HFpEF risk factors alone appears insufficient to explain the observed PAWP elevation. One additional contributor appears to be enhanced diastolic ventricular interaction from the right heart impacting LV filling as suggested by the flat PAWP waveforms without prominent v waves, evidence of pericardial restraint during exercise, and PH therapy induced reduction in both RA and PAWP in group 1 PH + high PAWP. There may also be pulmonary venous remodeling that could contribute to modest overestimation of LA pressure when using PAWP.30 Indeed, the observed acute and chronic reduction in PAWP with pulmonary vasodilators in group 1 PH + high PAWP might support this hypothesis as a reflection of pulmonary venodilation.31 Collectively, it seems likely that multiple independent pathophysiological mechanisms may variably contribute to the modest PAWP elevation in some patients with group 1 PH (Figure 4).
    Additionally, given that the group 1 PH + high PAWP group had the worst overall PA pressure elevation due to the superimposed impact of an increased PAWP on group 1 PH related PVR elevation, and the correlations observed between PAWP and PA pressure even in group 1 PH, there is a need for further study of therapeutic approaches to lower PAWP in these patients. Some patients with group 1 PH may develop LV cardiomyocyte atrophy from chronic LV underfilling32, which could contribute to elevated risk of LV chamber non-compliance that may manifest following pulmonary vasodilator induced improved LV filling or acquisition of superimposed HFpEF risk factors.33–35 There was notably a high burden of excess adiposity in group 1 PH + high PAWP, and with sodium glucose co-transporter 2 inhibitors (SGLT2i) and incretins demonstrating clear adipose reducing effects with associated heart failure benefits in patients with overt HFpEF36–39, further study of the added benefit of these therapies on top of pulmonary vasodilators in group 1 PH + high PAWP is needed.
    With SGLT2i, there is clear evidence of adipose tissue regression36,37,40 with associated resting and dynamic PAWP reduction36,41–44, with improved dynamic RV-PA coupling17,42. The effect of incretin therapy on PAWP remains unknown, but weight loss has been associated with improved PAWP and PH,45 and clinical trials of incretins in overt HFpEF have demonstrated improvement in natriuretic peptides, myocardial remodeling and functional status.38,39,46–48 Therefore, although there is increasing interest and guideline emphasis in identifying occult PAWP abnormalities with provocation among group 1 PH with normal resting PAWP,5,8 the current data suggest that a clinically important subset of patients with group 1 PH have overt elevation in PAWP even at rest, emphasizing the need for integrated probabilistic evaluation of all available information including hemodynamics, imaging, comorbidities and potentially even metabolomics49 to establish a clinical diagnosis of PAH. As a result of being excluded from all prior trials of group 1 PH, there is a pressing need for randomized trials to determine the optimal treatment strategies in this unique group 1 PH subgroup with high PAWP at rest.
    The present analysis has multiple strengths, including multiple validation cohorts, multicenter representation, manual re-adjudication of hemodynamic tracings, comprehensive functional and phenotypic characterization, metabolomic assessment, and diagnostic adjudication of all cases by a PH specialist. The prevalence of PAWP elevation in group 1 PH is higher than previously reported,50 which may relate to greater prevalence of obesity, and manual re-adjudication of PAWP measurement in all cases, with findings replicated with 4 independent cohorts. Although our approach to classify patients as having group 1 PH or Cpc PH HFpEF was consistent with current recommendations to utilize all available information in a probabilistic manner to make a binary decision regarding the most likely diagnosis, this process is inherently subjective and there is currently no existing operational framework to guide application. We therefore considered a series of clinical and hemodynamic criteria (table 1, and details in the online supplement) in each case through detailed chart/medical history review with manual hemodynamic tracing adjudication by a PH physician, and made a probabilistic diagnosis in each individual case of the presence of either group 1 PH or Cpc PH HFpEF. Although this approach is subjective by definition, it is consistent with contemporary clinical practice enhancing generalizability. We also tested concordance statistics between independent PH experts and demonstrated good diagnostic agreement for our adjudication process. However, it must be acknowledged that there is currently no objective ‘gold-standard’ to fully separate group 1 PH in the presence of risk factors for HFpEF from Cpc HFpEF, and we cannot exclude the possibility that some of the patients with group 1 PH + high PAWP may truly have a diagnosis of Cpc PH, and in many ways this is a key finding of this study. Future efforts are therefore needed to develop and test more objective multivariable based clinical criteria to separate group 1 PH from Cpc PH HFpEF without reliance on current binary hemodynamic thresholds alone. Acknowledging these limitations, the finding of large differences between group 1 PH + high PAWP and Cpc PH (and large similarities between group 1 PH with high or normal PAWP) using unbiased metabolomics, adjudication agreement by independent PH experts, and validation of our results in three independent cohorts, supports the likelihood that most of the study patients with group 1 PH + high PAWP had true group 1 PH as the primary pathology.

    Conclusion

    Although a resting PAWP>15 mm Hg is commonly used to implicate left heart failure as the cause of pulmonary hypertension, about 1 in 5 patients with adjudicated group 1 PH have elevated resting PAWP, which appears related in part to modest left heart diastolic dysfunction that is less severe than overt Cpc PH HFpEF. These patients display metabolomic signatures more typical of group 1 PH than Cpc PH, suggesting that they be classified as precapillary predominant PH, but the pathophysiologic differences present underline the priority for dedicated trials to guide treatment moving forward.

    Supplementary Material

    Expanded Methods
    Tables S1–S10
    Figures S1–S6
    Excel File S1
    Reference 51
    AHDr. Anna Hemnes

    Dr. Anna Hemnes

    Pulmonology & Critical Care · Pulmonary Vascular Medicine
    Vanderbilt Health
    annahemnes